ai-podcast-creation
Pass
Audited by Gen Agent Trust Hub on Feb 25, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill provides an installation command that fetches a shell script from the vendor's official domain (https://cli.inference.sh) and executes it directly. This is the intended method for installing the platform's 'infsh' command-line tool.
- [COMMAND_EXECUTION]: The skill requires the ability to execute commands starting with 'infsh' via a bash shell. This capability is used to run various AI models for audio processing, narration, and script generation.
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection because it processes external, potentially untrusted data (like document extracts or user-provided text) which is then passed into AI models as input. Evidence: 1. Ingestion points: The '--input' parameter in bash commands (e.g., in SKILL.md). 2. Boundary markers: No specific delimiters or safety instructions are used to separate untrusted content from the system prompt in the examples. 3. Capability inventory: The skill can execute subprocesses through the 'infsh' CLI and write output to files. 4. Sanitization: There is no evidence of input validation or escaping for the data processed by the models.
Audit Metadata